lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <649f0ea3bfae4_11e685294f2@dwillia2-xfh.jf.intel.com.notmuch>
Date:   Fri, 30 Jun 2023 10:19:31 -0700
From:   Dan Williams <dan.j.williams@...el.com>
To:     "Wilczynski, Michal" <michal.wilczynski@...el.com>,
        Dan Williams <dan.j.williams@...el.com>,
        <linux-acpi@...r.kernel.org>
CC:     <rafael@...nel.org>, <vishal.l.verma@...el.com>, <lenb@...nel.org>,
        <dave.jiang@...el.com>, <ira.weiny@...el.com>,
        <rui.zhang@...el.com>, <linux-kernel@...r.kernel.org>,
        <nvdimm@...ts.linux.dev>,
        "Rafael J . Wysocki" <rafael.j.wysocki@...el.com>
Subject: Re: [PATCH v5 09/10] acpi/nfit: Move handler installing logic to
 driver

Wilczynski, Michal wrote:
> 
> 
> On 6/29/2023 10:54 PM, Dan Williams wrote:
> > Michal Wilczynski wrote:
> >> Currently logic for installing notifications from ACPI devices is
> >> implemented using notify callback in struct acpi_driver. Preparations
> >> are being made to replace acpi_driver with more generic struct
> >> platform_driver, which doesn't contain notify callback. Furthermore
> >> as of now handlers are being called indirectly through
> >> acpi_notify_device(), which decreases performance.
> >>
> >> Call acpi_dev_install_notify_handler() at the end of .add() callback.
> >> Call acpi_dev_remove_notify_handler() at the beginning of .remove()
> >> callback. Change arguments passed to the notify function to match with
> >> what's required by acpi_install_notify_handler(). Remove .notify
> >> callback initialization in acpi_driver.
> >>
> >> Suggested-by: Rafael J. Wysocki <rafael.j.wysocki@...el.com>
> >> Signed-off-by: Michal Wilczynski <michal.wilczynski@...el.com>
> >> ---
> >>  drivers/acpi/nfit/core.c | 24 ++++++++++++++++++------
> >>  1 file changed, 18 insertions(+), 6 deletions(-)
> >>
> >> diff --git a/drivers/acpi/nfit/core.c b/drivers/acpi/nfit/core.c
> >> index 95930e9d776c..a281bdfee8a0 100644
> >> --- a/drivers/acpi/nfit/core.c
> >> +++ b/drivers/acpi/nfit/core.c
> >> @@ -3312,11 +3312,13 @@ void acpi_nfit_shutdown(void *data)
> >>  }
> >>  EXPORT_SYMBOL_GPL(acpi_nfit_shutdown);
> >>  
> >> -static void acpi_nfit_notify(struct acpi_device *adev, u32 event)
> >> +static void acpi_nfit_notify(acpi_handle handle, u32 event, void *data)
> >>  {
> >> -	device_lock(&adev->dev);
> >> -	__acpi_nfit_notify(&adev->dev, adev->handle, event);
> >> -	device_unlock(&adev->dev);
> >> +	struct acpi_device *device = data;
> >> +
> >> +	device_lock(&device->dev);
> >> +	__acpi_nfit_notify(&device->dev, handle, event);
> >> +	device_unlock(&device->dev);
> >>  }
> >>  
> >>  static int acpi_nfit_add(struct acpi_device *adev)
> >> @@ -3375,12 +3377,23 @@ static int acpi_nfit_add(struct acpi_device *adev)
> >>  
> >>  	if (rc)
> >>  		return rc;
> >> -	return devm_add_action_or_reset(dev, acpi_nfit_shutdown, acpi_desc);
> >> +
> >> +	rc = devm_add_action_or_reset(dev, acpi_nfit_shutdown, acpi_desc);
> >> +	if (rc)
> >> +		return rc;
> >> +
> >> +	return acpi_dev_install_notify_handler(adev,
> >> +					       ACPI_DEVICE_NOTIFY,
> >> +					       acpi_nfit_notify);
> >>  }
> >>  
> >>  static void acpi_nfit_remove(struct acpi_device *adev)
> >>  {
> >>  	/* see acpi_nfit_unregister */
> >> +
> >> +	acpi_dev_remove_notify_handler(adev,
> >> +				       ACPI_DEVICE_NOTIFY,
> >> +				       acpi_nfit_notify);
> > Please use devm to trigger this release rather than making
> > acpi_nfit_remove() contain any logic.
> 
> I think adding separate devm action to remove event handler is not
> necessary. I'll put the removal in the beggining of acpi_nfit_shutdown() if you
> don't object.

How do you plan to handle an acpi_dev_install_notify_handler() failure?
acpi_nfit_shutdown() will need to have extra logic to know that it can
skip acpi_dev_remove_notify_handler() in some cases and not other..
Maybe it is ok to remove a handler that was never installed, but I would
rather not go look that up. A devm callback for
acpi_dev_remove_notify_handler() avoids that.

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ